DESCRIPTION: (Applicant's Abstract) Transdermal Nicotine Replacement (TNR) is the best current addictions medication, doubling smoking cessation rates compared to placebo. Yet, the mechanism underlying its effects is still uncertain. Moreover, TNR typically fails: 79% of TNR patients relapse, two-thirds while still receiving nicotine. We propose to examine the mechanisms for TNR's success -- and its failures. TNR is posited to prevent relapse by blunting unconditioned withdrawal and craving, particularly early in cessation. TNR's effect on craving -- and its inadequacy in preventing relapse -- is reflected in our finding that smokers on TNR often relapse in the complete absence of subjective craving. We hypothesize that when craving is blunted by TNR, other factors such as conditioned responses to smoking cues, intrusive thoughts about smoking, and behavioral habits become important causes of relapse. We propose to assess these factors in the first intensive study of how agonist therapy affects the relapse process. To avoid relying on retrospective recall, we have developed a novel method using palm-top computers to obtain near-real-time reports of events in subjects' natural environment. We have demonstrated the feasibility and validity of the method for studying relapse, and shown that both global self-reports of smoking patterns and later recall of initial relapse episodes are inaccurate. In an amended (scaled down) design, we will use this method to study relapse in 353 smokers randomized, double-blind, to high-dose nicotine replacement (35 mg for 3 weeks, 21 mg for 2 weeks) or placebo. We will over-sample African-Americans to achieve 33% representation. Subjects will monitor ad lib baseline smoking for 1 week. The Electronic Diary (ED) will assess situational and affective antecedents of smoking, which relate to idiosyncratic conditioned associations. After quitting, subjects self-monitor for 6 weeks, recording temptation and lapse episodes for 5 weeks of active TNR and 1 week after TNR. Throughout, the ED prompts subjects for randomly-scheduled base-rate assessments (4/day). Withdrawal and craving are monitored several times daily. Once daily, ED assesses withdrawal-related changes in cognitive performance, experience of intrusive thoughts, behavioral automatisms and smoking expectancies. Our analytic approach emphasizes random-effects regression methods that allow for multiple observations per subject. The models include nicotine dependence and time since cessation as important moderating variables. This is not a treatment efficacy study; our unique focus is on fine-grained analysis of the relapse process, and how it is affected by pharmacologic treatment with TNR. By revealing why and how TNR fails, the findings will assist in development of improved behavioral strategies to supplement TNR; the results may also be relevant to medications development (particularly agonist drugs) for other drug addictions.